CIS 280 Data Visualization and Communication

Ch 6: Visualizing Amounts

---

Plots

1. Bar plots
   1. Single category bar plots
   2. Grouped / Stacked bar plots
2. Dot plots & Heatmaps

Use Cases

1. To count the number of times each category appears for a discrete data
2. To breakdown the total of some continuous variable by discrete variable

When to Use Each Type

1. If we have a single category to count or one set of continuous values to sum, we can use a bar plot.
2. If we have two or more sets of categories for which we want to show amounts, we can group or stack the bars.
3. If we have two or more sets of categories for which we want to show amounts on an x-y axis, we can group or dot plots.
4. If we want to see amount patterns in color

import plotly.express as px
import pandas as pd

from IPython.display import Image
from IPython.display import IFrame

6.1 Bar Plots

• Bar is the shape in our mapping elements - position, size, shape, color
• Visually, we compare the area of the bars
  • The width is the same for all bars - often do not set
  • The value or count determines the height - scaled from the data
• One dimension is the object being counted
  • One dimension must be qualitative.
  • A bar plot with only a continuous variable is a histogram
• The other dimension is the count or total
  • We often don't pass both X & Y arguments

# Table 2.1
IFrame(src='https://clauswilke.com/dataviz/aesthetic-mapping.html#tab:basic-data-types',width=1000,height=800)

df = px.data.tips()
df

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No	Sun	Dinner	2
1	10.34	1.66	Male	No	Sun	Dinner	3
2	21.01	3.50	Male	No	Sun	Dinner	3
3	23.68	3.31	Male	No	Sun	Dinner	2
4	24.59	3.61	Female	No	Sun	Dinner	4
...	...	...	...	...	...	...	...
239	29.03	5.92	Male	No	Sat	Dinner	3
240	27.18	2.00	Female	Yes	Sat	Dinner	2
241	22.67	2.00	Male	Yes	Sat	Dinner	2
242	17.82	1.75	Male	No	Sat	Dinner	2
243	18.78	3.00	Female	No	Thur	Dinner	2

244 rows × 7 columns

# counting qualitative data
px.bar(df,x='sex')

# total tip amount by sex
px.bar(df,y='sex')

# total tip amount by sex - note each slice represents the record value
px.bar(df,x='sex',y='tip')

# total tip amount by sex
px.bar(df,y='sex',x='tip')

# This is incorrect. Plot a single continuous variable with a histgram
px.bar(df,y='tip')

# Use a histogram to look at the distibution of continuous data
px.histogram(df,x='tip')

# Use a scatter plot for two continuous variables 
px.bar(df,y='tip',x='total_bill')

'''
Practice Exercise: Use the tips dataset to make two plots.

1. Make a bar plot that counts the number occurrences of a discrete variable.
2. Make a bar plot that breaks down a continuous variable by one discrete variable.
'''

df

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No	Sun	Dinner	2
1	10.34	1.66	Male	No	Sun	Dinner	3
2	21.01	3.50	Male	No	Sun	Dinner	3
3	23.68	3.31	Male	No	Sun	Dinner	2
4	24.59	3.61	Female	No	Sun	Dinner	4
...	...	...	...	...	...	...	...
239	29.03	5.92	Male	No	Sat	Dinner	3
240	27.18	2.00	Female	Yes	Sat	Dinner	2
241	22.67	2.00	Male	Yes	Sat	Dinner	2
242	17.82	1.75	Male	No	Sat	Dinner	2
243	18.78	3.00	Female	No	Thur	Dinner	2

244 rows × 7 columns

# Answer here

6.2 Grouped & Stacked Bar Plots

If we have two or more sets of categories for which we want to show amounts, we can group or stack the bars.

df

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No	Sun	Dinner	2
1	10.34	1.66	Male	No	Sun	Dinner	3
2	21.01	3.50	Male	No	Sun	Dinner	3
3	23.68	3.31	Male	No	Sun	Dinner	2
4	24.59	3.61	Female	No	Sun	Dinner	4
...	...	...	...	...	...	...	...
239	29.03	5.92	Male	No	Sat	Dinner	3
240	27.18	2.00	Female	Yes	Sat	Dinner	2
241	22.67	2.00	Male	Yes	Sat	Dinner	2
242	17.82	1.75	Male	No	Sat	Dinner	2
243	18.78	3.00	Female	No	Thur	Dinner	2

244 rows × 7 columns

# Map a second set of values to the color property
# Default is stacked bars
px.bar(df,x='sex',color='day')

# Add barmode to get a grouped plot
px.bar(df,x='sex',color='day',barmode='group')

# Breakdown a continuous variable by 2 qualitative variables
px.bar(df,x='sex',y='tip',color='day',barmode='group')

'''
Practice Exercise: Use the tips dataset to make one plot.

1. Make a bar plot that breaks down a total_bill by two discrete variables.
'''

df.head()

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No	Sun	Dinner	2
1	10.34	1.66	Male	No	Sun	Dinner	3
2	21.01	3.50	Male	No	Sun	Dinner	3
3	23.68	3.31	Male	No	Sun	Dinner	2
4	24.59	3.61	Female	No	Sun	Dinner	4

# Answer here

Aggregating Categorical Data

Image(url='https://clauswilke.com/dataviz/visualizing_amounts_files/figure-html/titanic-passengers-by-class-sex-1.png',width=600)

df

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No	Sun	Dinner	2
1	10.34	1.66	Male	No	Sun	Dinner	3
2	21.01	3.50	Male	No	Sun	Dinner	3
3	23.68	3.31	Male	No	Sun	Dinner	2
4	24.59	3.61	Female	No	Sun	Dinner	4
...	...	...	...	...	...	...	...
239	29.03	5.92	Male	No	Sat	Dinner	3
240	27.18	2.00	Female	Yes	Sat	Dinner	2
241	22.67	2.00	Male	Yes	Sat	Dinner	2
242	17.82	1.75	Male	No	Sat	Dinner	2
243	18.78	3.00	Female	No	Thur	Dinner	2

244 rows × 7 columns

# Get counts for all combinations as a series
df[['smoker','sex']].value_counts()

smoker  sex   
No      Male      97
Yes     Male      60
No      Female    54
Yes     Female    33
dtype: int64

# Get counts for all combinations as a dataframe
df[['smoker','sex']].value_counts().to_frame().reset_index()

	smoker	sex	0
0	No	Male	97
1	Yes	Male	60
2	No	Female	54
3	Yes	Female	33

df_agg = df[['smoker','sex']].value_counts().to_frame().reset_index().rename(columns={0:'count'})
df_agg

	smoker	sex	count
0	No	Male	97
1	Yes	Male	60
2	No	Female	54
3	Yes	Female	33

px.bar(df_agg,
       x='sex',
       y='count',
       color='smoker')

px.bar(df,
       x='sex',
       color='smoker')

6.3 Dot Plot

1. Dot plots are scatter plots with one categorical axis (Y) and one continuous axis (X).
2. Instead of using bars, we can also place dots at the location where the corresponding bar would end.
3. They can be used to show changes between two (or more) points in time or between two (or more) conditions.

Image(url='https://drive.google.com/uc?id=1aNIYjsmxpWPZEYHdPMhckRh2fq6_auN1')

df

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No	Sun	Dinner	2
1	10.34	1.66	Male	No	Sun	Dinner	3
2	21.01	3.50	Male	No	Sun	Dinner	3
3	23.68	3.31	Male	No	Sun	Dinner	2
4	24.59	3.61	Female	No	Sun	Dinner	4
...	...	...	...	...	...	...	...
239	29.03	5.92	Male	No	Sat	Dinner	3
240	27.18	2.00	Female	Yes	Sat	Dinner	2
241	22.67	2.00	Male	Yes	Sat	Dinner	2
242	17.82	1.75	Male	No	Sat	Dinner	2
243	18.78	3.00	Female	No	Thur	Dinner	2

244 rows × 7 columns

df.day.value_counts()

Sat     87
Sun     76
Thur    62
Fri     19
Name: day, dtype: int64

# Counting bar chart with 1 categorical and 1 continuous variable
px.bar(df,x='day',y='tip')

# We can't just convert a bar to a scatter - we'll just show each record
px.scatter(df,y='day',x='tip')

Aggregating Mixed Data

Pandas Group By
Group a dataframe by a categorical variable and compute some aggregate function on a continuous variable

• https://pandas.pydata.org/pandas-docs/dev/reference/api/pandas.DataFrame.groupby.html
• Split/Apply/Combine - split the df into groups, apply a function to each group, combine groups
• Similar to looping, but easier

# Group rows by by sex, then apply a count function to the tip, then combine into one structure
df.groupby(by='sex')['tip'].sum()

sex
Female    246.51
Male      485.07
Name: tip, dtype: float64

# Can have multiple categories, compute a function across multiple continuous variables, and select one of many functions
df.groupby(by=['day','sex'])['tip','total_bill'].mean()

/var/folders/kh/plns67p91gv9v06gw7hpyxg80000gn/T/ipykernel_93291/1964103464.py:2: FutureWarning:

Indexing with multiple keys (implicitly converted to a tuple of keys) will be deprecated, use a list instead.

		tip	total_bill
day	sex
Fri	Female	2.781111	14.145556
	Male	2.693000	19.857000
Sat	Female	2.801786	19.680357
	Male	3.083898	20.802542
Sun	Female	3.367222	19.872222
	Male	3.220345	21.887241
Thur	Female	2.575625	16.715312
	Male	2.980333	18.714667

df_agg = df.groupby(by=['day','smoker'])['tip'].sum().to_frame().reset_index()

df_agg

	day	smoker	tip
0	Fri	No	11.25
1	Fri	Yes	40.71
2	Sat	No	139.63
3	Sat	Yes	120.77
4	Sun	No	180.57
5	Sun	Yes	66.82
6	Thur	No	120.32
7	Thur	Yes	51.51

# Use aggregated data to plot the combined values
px.scatter(df_agg,
           y='day',
           x='tip',
           color='smoker')

'''
Practice Exercise: Use groupby to aggregate.

1. Make a bar plot that breaks down a total_bill by two discrete variables.
'''

df.head()

	total_bill	tip	sex	smoker	day	time	size
0	16.99	1.01	Female	No	Sun	Dinner	2
1	10.34	1.66	Male	No	Sun	Dinner	3
2	21.01	3.50	Male	No	Sun	Dinner	3
3	23.68	3.31	Male	No	Sun	Dinner	2
4	24.59	3.61	Female	No	Sun	Dinner	4

# Answer here

6.3+ Heatmap

With px.imshow, each value of the data frame is represented as a heatmap pixel

# load stocks with continuous data
df = px.data.stocks()
df

	date	GOOG	AAPL	AMZN	FB	NFLX	MSFT
0	2018-01-01	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000
1	2018-01-08	1.018172	1.011943	1.061881	0.959968	1.053526	1.015988
2	2018-01-15	1.032008	1.019771	1.053240	0.970243	1.049860	1.020524
3	2018-01-22	1.066783	0.980057	1.140676	1.016858	1.307681	1.066561
4	2018-01-29	1.008773	0.917143	1.163374	1.018357	1.273537	1.040708
...	...	...	...	...	...	...	...
100	2019-12-02	1.216280	1.546914	1.425061	1.075997	1.463641	1.720717
101	2019-12-09	1.222821	1.572286	1.432660	1.038855	1.421496	1.752239
102	2019-12-16	1.224418	1.596800	1.453455	1.104094	1.604362	1.784896
103	2019-12-23	1.226504	1.656000	1.521226	1.113728	1.567170	1.802472
104	2019-12-30	1.213014	1.678000	1.503360	1.098475	1.540883	1.788185

105 rows × 7 columns

# Each value is a color bar - raw values
px.imshow(df,width=700,height=700)

# Use pandas correlation function
df.corr()

	GOOG	AAPL	AMZN	FB	NFLX	MSFT
GOOG	1.000000	0.833629	0.556702	0.633169	0.140254	0.747029
AAPL	0.833629	1.000000	0.560877	0.493498	0.049519	0.786771
AMZN	0.556702	0.560877	1.000000	0.341430	0.619946	0.660896
FB	0.633169	0.493498	0.341430	1.000000	0.265663	0.472227
NFLX	0.140254	0.049519	0.619946	0.265663	1.000000	0.079532
MSFT	0.747029	0.786771	0.660896	0.472227	0.079532	1.000000

# Matrix of values
px.imshow(df.corr())

Use imshow to plot 3 color images

import numpy as np

colors = np.random.random(9*16).reshape(9,16)
colors

array([[0.00661908, 0.38889081, 0.82299527, 0.29400284, 0.51999436,
        0.92975834, 0.16530968, 0.87993949, 0.62508173, 0.68918127,
        0.75492581, 0.44298281, 0.72560364, 0.71857954, 0.42530702,
        0.38166019],
       [0.55843956, 0.09340349, 0.06804757, 0.56337219, 0.17338547,
        0.15427628, 0.13325399, 0.72909297, 0.76820082, 0.97933499,
        0.95486782, 0.29918547, 0.92712032, 0.24453814, 0.15759402,
        0.94983585],
       [0.46073902, 0.47604746, 0.26661154, 0.4310107 , 0.8657953 ,
        0.90104762, 0.43565712, 0.02279515, 0.27638027, 0.46970307,
        0.15344466, 0.78965919, 0.19433078, 0.30735017, 0.82680199,
        0.60376475],
       [0.6945631 , 0.8436158 , 0.35036677, 0.36266975, 0.446323  ,
        0.01392668, 0.01973868, 0.64562098, 0.8615941 , 0.03255129,
        0.74675502, 0.16225547, 0.0091061 , 0.70497683, 0.11507202,
        0.23814179],
       [0.52843522, 0.40524668, 0.05614435, 0.26303703, 0.31930379,
        0.83882764, 0.30817174, 0.4568098 , 0.9796798 , 0.68190925,
        0.5803533 , 0.91400869, 0.18450815, 0.1643243 , 0.09157274,
        0.14142348],
       [0.11454467, 0.22625487, 0.34165394, 0.02946669, 0.36934826,
        0.19494686, 0.46765759, 0.20477195, 0.49988452, 0.93330036,
        0.9748635 , 0.94530959, 0.00620233, 0.37153224, 0.83672787,
        0.63673004],
       [0.96753576, 0.96360787, 0.65414457, 0.79004722, 0.80943304,
        0.61042928, 0.64677509, 0.55435694, 0.20431713, 0.09773841,
        0.3581671 , 0.5596828 , 0.71933434, 0.36892599, 0.65870409,
        0.03320889],
       [0.21118429, 0.31646121, 0.99402624, 0.46530278, 0.46162795,
        0.18405642, 0.76823483, 0.69429339, 0.13512095, 0.81381762,
        0.37039597, 0.10622007, 0.07123317, 0.35011568, 0.43613387,
        0.17248956],
       [0.27848561, 0.67316211, 0.92892431, 0.19603397, 0.05025232,
        0.43824633, 0.92738859, 0.97291609, 0.92169445, 0.45274421,
        0.49993208, 0.61337845, 0.38342769, 0.17913624, 0.76408922,
        0.0781835 ]])

px.imshow(colors)